At the European Central Bank (ECB) annual conference in Sintra this year, discussion was dominated by how artificial intelligence (AI) could transform the global economy and what that means for preserving financial stability. Central bankers, IMF officials and private-sector experts debated AI's potential effects across markets, lending, security policy and energy demand.
Participants agreed that AI could produce wide-ranging and hard-to-quantify consequences. The technology's impact may be profound but ambiguous: it can boost efficiency and growth while simultaneously creating new systemic vulnerabilities.
Dual-sided effects: growth driver and source of fragility
Torsten Slok of Apollo Global Management warned that AI's consequences are two-sided: whether the technology performs better or worse than expected, it will significantly affect financial stability. Slok estimated that the substantial capital flowing into the AI sector has itself raised U.S. GDP by about one percentage point.
The Bank for International Settlements (BIS) likened the current investment wave to some of history's largest speculative episodes, citing parallels with the British railway mania of the 1840s and the dot-com bubble.
Bubble and manipulation risks
Speakers cautioned that AI could rapidly inflate and then puncture asset bubbles while profiting from both phases, effectively enabling forms of market manipulation. Itay Goldstein of the University of Pennsylvania warned that algorithmic trading could steer prices onto manipulative trajectories, producing artificial rallies and sudden crashes.
Lending, opacity and inequality
AI can enable more precise credit-risk assessment and widen access to financing for previously excluded borrowers. However, Tobias Adrian of the International Monetary Fund (IMF) highlighted that machine-learning credit models often operate as "black boxes," complicating supervisory oversight and the explanation of decisions to authorities.
Moreover, rising costs for cyber defense tied to AI systems may exacerbate gaps between richer and poorer firms and countries.
Supervisory responses and backstops
Sarah Breeden, Deputy Governor of the Bank of England, proposed a protective framework modeled on deposit insurance used in bank failures. Such mechanisms would aim to limit systemic fallout if AI-related disruptions precipitate severe stress in financial institutions.
The paradox: success can also be risky
Speakers noted a paradoxical central risk: if AI meets optimistic expectations for efficiency gains, machines could displace large numbers of workers, leading to higher unemployment, lower incomes and ultimately economic contraction. Conversely, if the technology underdelivers, the huge capital invested in the sector may fail to generate expected returns.
Tiff Macklem, Governor of the Bank of Canada, recalled that while the internet ultimately created whole new industries, the bursting of the dot-com bubble could not be avoided.
Conclusion
The Sintra conference participants concluded that AI introduces novel risks to the financial system: it can amplify speculation, complicate supervision and deepen socio-economic disparities. Strengthening regulatory tools and increasing transparency will be central to preserving financial stability as AI adoption accelerates.
This article summarizes views expressed at the conference and does not constitute investment advice.
